biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bugfix] Fix Typo from NVFP4 Refactor (#31977)

Browse Source 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
This commit is contained in:
Robert Shaw
2026-01-08 19:18:50 -05:00
committed by GitHub
parent 6ebe34d6fa
commit 0fa8dd24d2
6 changed files with 30 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tests/evals/gsm8k/configs/moe-refactor-dp-ep/Qwen3-30B-A3B-NvFp4-CT-fi-cutedsl-deepep-ll.yaml Normal file
View File

@@ -0,0 +1,8 @@
model_name: "RedHatAI/Qwen3-30B-A3B-NVFP4"
accuracy_threshold: 0.88
num_questions: 1319
num_fewshot: 5
server_args: "--enforce-eager --max-model-len 8192 --data-parallel-size 2 --enable-expert-parallel --all2all-backend deepep_low_latency"
env:
VLLM_USE_FLASHINFER_MOE_FP4: "1"
VLLM_FLASHINFER_MOE_BACKEND: "masked_gemm"

8
tests/evals/gsm8k/configs/moe-refactor-dp-ep/Qwen3-30B-A3B-NvFp4-ModelOpt-fi-cutedsl-deepep-ll.yaml Normal file
View File

@@ -0,0 +1,8 @@
model_name: "nvidia/Qwen3-30B-A3B-NVFP4"
accuracy_threshold: 0.88
num_questions: 1319
num_fewshot: 5
server_args: "--enforce-eager --max-model-len 8192 --data-parallel-size 2 --enable-expert-parallel --all2all-backend deepep_low_latency"
env:
VLLM_USE_FLASHINFER_MOE_FP4: "1"
VLLM_FLASHINFER_MOE_BACKEND: "masked_gemm"

8
tests/evals/gsm8k/configs/moe-refactor-dp-ep/Qwen3-30B-A3B-NvFp4-ModelOpt-fi-cutlass.yaml Normal file
View File

@@ -0,0 +1,8 @@
model_name: "nvidia/Qwen3-30B-A3B-NVFP4"
accuracy_threshold: 0.88
num_questions: 1319
num_fewshot: 5
server_args: "--enforce-eager --max-model-len 8192 --data-parallel-size 2 --enable-expert-parallel"
env:
VLLM_USE_FLASHINFER_MOE_FP4: "1"
VLLM_FLASHINFER_MOE_BACKEND: "throughput"

6
tests/evals/gsm8k/configs/moe-refactor-dp-ep/config-b200.txt
View File

@@ -1,8 +1,10 @@
Qwen3-30B-A3B-NvFp4-CT-fi-cutedsl-deepep-ll.yaml
Qwen3-30B-A3B-NvFp4-ModelOpt-fi-cutedsl-deepep-ll.yaml
Qwen3-30B-A3B-NvFp4-CT-fi-cutlass.yaml
Qwen3-30B-A3B-NvFp4-ModelOpt-fi-cutlass.yaml
Qwen3-30B-A3B-Fp8-AutoFp8-deepgemm-deepep-ht.yaml
Qwen3-30B-A3B-Fp8-AutoFp8-deepgemm-deepep-ll.yaml
Qwen3-30B-A3B-Fp8-AutoFp8-deepgemm.yaml
Qwen3-30B-A3B-Fp8-CT-Block-deepgemm-deepep-ht.yaml
Qwen3-30B-A3B-Fp8-CT-Block-deepgemm-deepep-ll.yaml
Qwen3-30B-A3B-Fp8-CT-Block-deepgemm.yaml
Qwen3-30B-A3B-NvFp4-CT-fi-cutlass.yaml
Qwen3-30B-A3B-NvFp4-ModelOpt-fi-cutlass-fi-a2av.yaml

1
tests/evals/gsm8k/configs/moe-refactor/config-test.txt Normal file
View File

@@ -0,0 +1 @@
Qwen3-30B-A3B-NvFp4-CT-marlin.yaml

2
vllm/model_executor/layers/quantization/utils/flashinfer_fp4_moe.py
View File

@@ -456,7 +456,7 @@ def prepare_nvfp4_moe_layer_for_fi_or_cutlass(
NvFp4MoeBackend.VLLM_CUTLASS,
NvFp4MoeBackend.FLASHINFER_CUTLASS,
NvFp4MoeBackend.FLASHINFER_TRTLLM,
NvFp4MoeBackend.FLASHINFER_TRTLLM,
NvFp4MoeBackend.FLASHINFER_CUTEDSL,
]
# Reorder [w1, w3] to [w3, w1] for FI NVFP4 MoE kernels.